morning Cyrill,

Well, the most elegant way to convert back and forth between numerical (ASCII-like) values and text is Martin Peach's string patch to the pd sources (search the PD-dev archives for "string type for pd"). If you don't feel like going that deep, you can use my [any2string] and [string2any] externals (in CVS, or from me at www.ling.uni-potsdam.de/~moocow/projects/pd). Alternatively, you could split symbols into single characters using [symbol2list] from zexy, and then map these to numbers however you want (e.g. with [index] from zexy, PDContainer, pool, [gfsm_alphabet] from gfsm, etc.) You can do character classification (and even substring classification) with python (in which case you don't really need the string or indexing stuff), or with a simple finite state transducer (gfsm again), although if you're just interested in differentiating between consonants & vowels, you're probably better off just using [route] here...

marmosets, Bryan

On 2007-05-11 21:33:29, Cyrill cduneau@yahoo.com appears to have written:

Hi list

I am willing to set up a patch that could convert letters to numerical values, so i could use converted text to send data to other patches (and then produce sound). My problem is that i want to work with hebrew, and i want pd to be able to differenciate between consonants and vowels, and maybe even to be able to recognize roots within the word. As I am quite a novice concerning pd, I have no clue about the most efficient way to convert letters into values, and i dont even know if it is possible within pd?

Any advice will be greatly appreciated.

Shalom

Cyrill